No, you are assuming so.
We don't know the algorithm in GMATPrep. In the old DOS version (PowerPrep), people figured out what problems were "not scored". In the newer versions, this is unknown.
Someone can miss 40% of the problems and score a 500, while someone else can miss 40% of the problems and score a 720. Luck may play a small role, but when you compound the 20% chance per problem over 37 or 41 problems, luck goes right out the door.
Don't over analyze these things. It's not worth the stress/worry.
gurpreetsingh
No I m not assuming this, just by chance they score same number of correct/incorrects and their level is also almost same. But there will be surely difference in the marks..
So getting questions wrong ONLY on incorrect group is a matter of luck, isnt it?